Metabase Automation

The classic Metabase REST API owns cards, dashboards, collections, permissions, and schema refresh. The versioned Agent API is the surface for headless semantic BI assistants and app-side AI workflows. The official MCP server (added in v60) and @metabase/cli (added in v62) serve AI clients and terminal scripting. Which release added what is in references/metabase-59-surface.md; before citing version-specific behavior, read the user's version from GET /api/session/properties (version), the current stable line from metabase.com/releases, and behavior from /docs/latest.

Quick Reference

Task	Path	Use When
Create/update questions and dashboards	Classic REST API + scripts/metabase_api.py	Standard content automation and incremental upserts
Promote content between environments	Remote Sync or serialization	Git-backed promotion, reviewable diffs, cross-environment moves
Build embedded customer analytics	Embedding + tenants + embedding permissions	Multi-tenant apps, customer portals, row-level isolation
Build an AI analytics app	Agent API	Versioned, semantic, app-side AI querying
Integrate Metabase with an AI coding agent	MCP server (v60+)	Claude, Cursor, VS Code — generate questions and dashboards via conversation
Govern AI access by group	Metabot AI governance (v61+, Pro/Enterprise)	Per-group controls, token limits, usage analytics
Refresh schema metadata	Database sync/rescan endpoints	New tables, changed columns, stale field values
Tune native SQL questions	Export-first + native query patterns	Stable automation without guessing request shapes

Decision Tree

Need to create or edit saved Metabase content?
  -> Use classic REST API (`card`, `dashboard`, `collection`).

Need repeatable dev -> prod promotion with reviewable diffs?
  -> Prefer Remote Sync or serialization before raw REST upserts.

Need embedded analytics for many customers or workspaces?
  -> Use embedding + tenants + embedding permissions.

Need an AI assistant to discover metrics/tables and construct queries?
  -> Use the versioned Agent API, not raw card CRUD.

Need to generate or edit questions/dashboards from an AI coding agent or terminal?
  -> Use the official Metabase MCP server (v60+, connects Claude/Cursor/VS Code).

Quick Start

Inputs (env vars)

• METABASE_URL (e.g., https://metabase.example.com)
• Preferred: METABASE_API_KEY
• Optional: METABASE_SESSION
• Fallback: METABASE_USERNAME + METABASE_PASSWORD

Sanity checks

python3 skills/universal/data-metabase/scripts/metabase_api.py health
python3 skills/universal/data-metabase/scripts/metabase_api.py whoami
python3 skills/universal/data-metabase/scripts/metabase_api.py list-databases
python3 skills/universal/data-metabase/scripts/metabase_api.py list-collections --tree

Live API documentation

Your Metabase instance serves OpenAPI docs at /api/docs (for example https://metabase.example.com/api/docs). Use this to confirm request shapes for your exact build before scripting bulk edits.

Workflow

1. Confirm API availability (GET /api/util/health).
2. Authenticate with an API key first, then fall back to session auth only if needed.
3. Discover IDs instead of hardcoding them across environments:
   • collection_id for save location
   • database id for dataset_query
   • table and field IDs if using MBQL
4. Choose the right surface:
   • classic REST API for cards/dashboards/admin
   • Agent API for semantic, AI-driven querying
   • Remote Sync or serialization for promotion workflows
5. Create/update a card:
   • prefer native SQL for stable automation
   • set display and visualization_settings explicitly
6. Create/update a dashboard, then place cards with explicit layout.
7. Refresh metadata if schema changed.
8. Validate by running/exporting results and re-opening exported JSON.

Mutation and Promotion Gate

Treat a Metabase content change as a small migration, especially when dashboard filters or permissions are involved:

1. Export current card/dashboard JSON for diagnosis and record stable entity IDs or environment-local lookup keys. Do not call that read response a restore artifact.
2. Resolve every database, collection, table, field, and group in the target environment; do not translate numeric IDs by position.
3. Produce the intended payload and a mutation inventory (create, update, archive, permission change). For each operation, record its supported inverse, exact writable fields, target identifiers, and dependencies. Inventory permissions, archive state, cards, dashboards, dashcard layout, and filter mappings separately.
4. Apply dependency order: collections and permissions -> cards -> dashboards -> dashcard layout/filter mappings.
5. Re-export the result and compare semantic fields, layout, permissions, and query output to the intended payload.
6. Before production mutation, rehearse the selected restore path in staging or a disposable collection. A JSON read response is evidence, not automatically a replayable write payload.
7. If validation fails, stop later mutations and execute only the proven path: endpoint-specific inverse writes; a Serialization export/import covering those entities; or a Remote Sync revision for self-contained synced content. Verify product/license support and coverage first: Serialization and Remote Sync omit users, groups, and permissions, while Remote Sync also omits table metadata, alerts, and subscriptions. Restore excluded state through its owning API or documented admin workflow, then re-export and rerun queries to prove recovery.

An upsert is idempotent only when its lookup key is unique and stable in that environment. If two candidates match, stop and surface the ambiguity rather than updating the first result.

Key Concepts

• UI "Question" == API card
• Chart configuration lives on the card as display + visualization_settings
• Most visualization keys are easiest to manage by exporting an existing card JSON, then editing that payload
• Remote Sync is for Git-backed promotion, not per-request runtime automation
• Agent API is a separate, versioned surface for semantic query construction and execution; it caps rows per response (read the cap from the Agent API docs), so always handle continuation_token pagination
• MCP server (v60+) connects Metabase to Claude, Cursor, VS Code, and other MCP clients with the same permissions as the authenticated user; enables dashboards-as-code from AI terminals; v62 adds SQL execution, collection creation, and interactive charts rendered in the AI client itself
• Official Metabase CLI (npm install -g @metabase/cli; check the CLI README for the minimum Metabase version, which differs for API-key and browser login; distinct from the classic REST/JAR admin surfaces) builds questions, dashboards, documents, and transforms from the terminal — see references/metabase-cli.md
• Metabot (v60+, open source) integrates with Slack; v61 adds per-group access controls, token limits, and usage analytics (Pro/Enterprise only for governance features)
• Modular embedding SDK is a paid-plan feature. Read its React, Node, and Metabase/SDK version-pair requirements from /docs/latest/embedding before scaffolding or upgrading an app. Breaking boundary: SDK 1.54.x and below use the legacy auth API (authProviderUri, backend returns { id }); later versions drop authProviderUri and the backend returns { jwt }
• Custom visualizations (v62+, paid plans) let you build React-based chart types via a plugin SDK; check the docs for which surfaces render them (embeds, subscriptions, and alerts did not at launch) before promising one there

Plans and Pricing

Plan tiering decides many designs: SSO, row and column security (sandboxing), the modular embedding SDK, custom visualizations, and Metabot/AI governance have been paid-plan features, not Open Source. Tier names and which features sit on which tier change, so confirm the feature on the metabase.com/pricing comparison before designing around it.

Do not quote prices from memory. Before putting a cost in a proposal, read metabase.com/pricing and record: base price and included users per tier, per-user price, whether the figure shown is the annual-billing or monthly-billing price (they differ), the hosted AI token rate versus bring-your-own-provider-key, and the transforms allowance and per-run price.

Guardrails

• Prefer Remote Sync or Metabase serialization for bulk, cross-environment promotion; use direct API for incremental upserts.
• Do not hardcode numeric IDs across environments when you can discover them or use entity IDs / synced content.
• Never commit METABASE_API_KEY, passwords, or session tokens.
• Prefer a dedicated, least-privileged automation account and collection.
• Treat internal admin notify endpoints separately from normal automation. They use server-side MB_API_KEY, not the user-facing METABASE_API_KEY.
• If a dashboard uses tabs, filters, or complex dashcard state, export first and preserve that structure instead of rebuilding from memory.

Navigation

Topic	File	Load when
Authentication (API key + fallback)	references/api-auth.md	Any API automation task
Reports (cards): create/edit patterns	references/reports-cards.md	Creating or updating saved questions
Dashboards and card placement	references/dashboards.md	Building or replicating dashboards
Charts and visualization_settings	references/charts-settings.md	Configuring chart display or axis settings
Agent API for semantic BI	references/agent-api.md	Building AI analytics apps or headless BI workflows
Embedding and integration	references/embedding-integration.md	Public links, signed JWT, or SDK embedding
Tenants, embedding permissions, routing	references/tenants-routing.md	Multi-tenant apps or customer portal isolation
Permissions and collections	references/permissions-collections.md	Group setup, sandboxing, or collection hierarchy
Native SQL query patterns	references/native-query-patterns.md	Template tags, field filters, caching in SQL cards
Remote Sync and promotion workflows	references/remote-sync.md	Git-backed content promotion or cross-env moves
Surface map and release anchors	references/metabase-59-surface.md	Confirming which API layer a request belongs to, or which release added a feature
CLI: mb API client and JAR admin commands	references/metabase-cli.md	Scripting content via mb, or server admin (migrate, dump-to-h2, serialization)

Assets

Template	File
Card spec skeleton	assets/card-spec.template.json
Dashboard spec skeleton	assets/dashboard-spec.template.json
Dashcard layout skeleton	assets/dashcards-layout.template.json
Embed JWT examples	assets/embed-jwt-example.md

Scripts

scripts/metabase_api.py is a dependency-free helper for auth, discovery, content CRUD, dashboard layout work, query execution, and schema refresh. Dashboard placement writes require explicit arrays and unique nonzero integer ids; malformed fetched lists stop the write because omitted placements are deleted. Run its offline regressions with python3 -m pytest scripts/test_metabase_api.py from this skill directory.

Examples:

# Print authenticated user (tries API key, then session)
python3 skills/universal/data-metabase/scripts/metabase_api.py whoami

# Discover IDs before scripting
python3 skills/universal/data-metabase/scripts/metabase_api.py list-collections --tree
python3 skills/universal/data-metabase/scripts/metabase_api.py list-databases
python3 skills/universal/data-metabase/scripts/metabase_api.py database-metadata --id 2
python3 skills/universal/data-metabase/scripts/metabase_api.py list-fields --database-id 2

# Export an existing card JSON (use as a template for visualization_settings)
python3 skills/universal/data-metabase/scripts/metabase_api.py export-card --id 123 --out card.json

# Export an existing dashboard JSON (use as a template for layout)
python3 skills/universal/data-metabase/scripts/metabase_api.py export-dashboard --id 5 --out dashboard.json

# Create/update a card from a JSON spec (see references/reports-cards.md)
python3 skills/universal/data-metabase/scripts/metabase_api.py upsert-card --spec card-spec.json

# Create/update a dashboard from a JSON spec
python3 skills/universal/data-metabase/scripts/metabase_api.py upsert-dashboard --spec dashboard-spec.json

# Add or update dashboard layout
python3 skills/universal/data-metabase/scripts/metabase_api.py add-dashcard --dashboard-id 5 --spec dashcard.json
python3 skills/universal/data-metabase/scripts/metabase_api.py update-dashcards --dashboard-id 5 --spec dashcards-layout.json

# Execute a query spec or export a saved card result
python3 skills/universal/data-metabase/scripts/metabase_api.py run-query --spec dataset-query.json
python3 skills/universal/data-metabase/scripts/metabase_api.py export-card-query --id 123 --format csv --out report.csv

# Refresh metadata after schema changes
python3 skills/universal/data-metabase/scripts/metabase_api.py sync-schema --id 2
python3 skills/universal/data-metabase/scripts/metabase_api.py rescan-values --id 2

Known Traps

• Treating Metabase as the semantic layer of record while metrics, joins, and business definitions still drift in dbt, SQLMesh, or raw SQL cards.
• Copying cards manually between environments and silently breaking references, permissions, or dashboard filter wiring because serialization or sync strategy was never defined.
• Reusing one question across dashboards with slightly different filter semantics and creating hidden KPI disagreements that look like product changes.
• Embedding tenant dashboards without a strict parameter contract, collection isolation model, and signed-embed verification path.
• Letting native SQL cards become the default authoring path for everything, then losing discoverability, lineage, and reuse across teams.
• Designing dashboard filters around visual convenience rather than stable field mappings, leading to broken linked filters after schema or model changes.

Common Anti-Patterns

• Treating dashboard screenshots or PDF exports as the decision artifact instead of versioned questions, documented metrics, and reproducible model definitions.
• Building one large "executive dashboard" that mixes raw exploratory cards, production KPIs, and operational monitoring in the same surface.
• Using personal collections as production distribution channels instead of curated shared collections with ownership and promotion rules.
• Hardcoding environment-specific table names, URLs, or card IDs into migration and automation workflows.
• Solving multi-tenancy with card duplication alone when embedding parameters, row-level source models, or separate collections would provide cleaner isolation.
• Using Metabase to compensate for unresolved source-model problems that should be fixed upstream in warehouse modeling or metric governance.

Related Skills

• ../data-analytics-engineering/SKILL.md for semantic modeling, metrics, and BI-ready warehouse design
• ../data-sql-optimization/SKILL.md for query tuning and database performance work
• ../software-backend/SKILL.md for application integration and service-layer implementation
• ../agents-mcp/SKILL.md when Metabase access should be wrapped as tools rather than ad hoc scripts
• ../agents-mcp/references/mcp-for-dwh.md for the LLM-via-MCP pattern catalog (Shape A managed-BI MCP applies directly to Metabase); enforcement layers (read-only role, query budgets, pseudonym boundary, audit log) apply regardless of shape

Learnings Loop

When prior decisions or pitfalls are relevant, consult learnings.consolidated.md if present; use learnings.md only for needed history or as the available fallback. Otherwise skip both.

After applying it, if you encountered a pattern worth remembering, a mistake worth preventing, or a domain fact that surprised you, append one dated bullet to learnings.md via agents-skills-feedback-loop/scripts/append_learning.py. Do not modify SKILL.md itself.